And What Are the Pitfalls in Gender Study?

William Petty's aphorism, "to measure is the first step to improve", succinctly summarizes the role of economic and social indicators, like those of income, health, nutrition and education, in mainstreaming gender. Between measurement and improvement, however, lies the stage of interpretation, and it is as well to be aware of possible pitfalls, mainly three: inappropriate indicators, use and interpretation. Consider the following examples:

Inappropriate indicators:

The use of life expectancy as a health indicator. In many developing countries, female mortality continues to be higher than male in younger age groups, despite biological reasons to believe that males are more susceptible. Thus mortality rates (or sex ratios) at specific age groups may be more appropriate indicators of gender discrimination in health.

Differences in male and female enrolment rates cannot automatically be interpreted as reflecting equal access. If girls have to miss school more often than boys or cannot give as much time as boys to their homework after school in order to do household reproductive work, this means less access even though enrolment rates are the same for males and females.

Use of anthropometric measures like height and weight to assess gender bias cannot pick up situations where discrimination is so severe and directed as to result in death (e.g. abortion of unwanted girl children), rather than causing a general neglect. A more appropriate indicator would be "sex ratio at birth".

Inappropriate use of indicators:

Anthropometric measurement norms below which a person is to be considered undernourished could be derived from the local population from amongst a well-fed group. If gender inequality already exists here, such norms render gender-selective under-nutrition unnoticed.

In assessing gender differences in food intake, if the difference in energy requirements between males and females is not taken into account, a higher energy intake by males may be wrongly construed as indicative of anti-female discrimination, just as workload of women, especially those involved in hard agricultural labour and heavy household work, may be underestimated, with disadvantages in energy intake-expenditure relationships being undetected.

Misinterpretation of indicators: Indicators convey differing meanings in specific social contexts. If information obtained from indicators is interpreted outside the context it was obtained from, it can be misleading. Consider the following:

Female mortality rates in South-east Asian countries, which were earlier found to be higher than for males, have now reduced. This could he interpreted as indicating the disappearance of a previous gender bias. A further exploration, however, reveals that rates of abortion of female foetuses in these regions have shown a steep increase over the same time period. Thus the gender bias still exists but is operating at a more elementary level.

In Bangladesh and Nepal, xerophthalmia, common in infants and pre-school children and resulting from Vitamin A deficiency, was found to afflict males almost twice as frequently as females. This could suggest either inherent male vulnerability or neglect. Instead, male favouritism, which results in weaned males being given "cultural superfood" while keeping females fully breast-fed, appears to be the source of male deprivation of Vitamin A.

Indicators are not intended to reveal the social processes that cause gender differences. Policy responses, when restricted to specific interventions in the areas revealed by indicators to show gender difference to the neglect of underlying determinants, address the symptoms not the causes. For example, data that reveal significantly higher female mortality than that of male may result in policies directed towards reducing female mortality by improving female health care and nutrition. Such higher mortality may, however, be an aspect of general discrimination, manifested in ways ranging from the economic to the cultural, which may prove to be more fundamental issues.

Indicators thus reflect "what is" without by themselves revealing "why it is". The interpretation of information thus requires local empirical grounding. Separate in-depth study into causality is essential for framing effective policies. They are thus necessary, but not sufficient, to devise and implement action towards gender equity. Appropriate indicators need to be used, together with appropriate controls and norms, plus careful interpretation within the right context, together with (case study) information on causality. We cannot cover here the pitfalls of all indicators thought useful to assess gender bias, but do wish to sound a note of caution of the ways in which gender indicators may prove to be deceptive for policy purposes.